Gaetan Lepage
459 lines
14 KiB
Nix
459 lines
14 KiB
Nix
{
|
|
lib,
|
|
buildPythonPackage,
|
|
fetchFromGitHub,
|
|
|
|
# build-system
|
|
hatchling,
|
|
|
|
# dependencies
|
|
aiosqlite,
|
|
appdirs,
|
|
chromadb,
|
|
click,
|
|
instructor,
|
|
json-repair,
|
|
json5,
|
|
jsonref,
|
|
litellm,
|
|
mcp,
|
|
openai,
|
|
opentelemetry-api,
|
|
opentelemetry-exporter-otlp-proto-http,
|
|
opentelemetry-sdk,
|
|
openpyxl,
|
|
pdfplumber,
|
|
portalocker,
|
|
pydantic,
|
|
pydantic-settings,
|
|
pyjwt,
|
|
python-dotenv,
|
|
regex,
|
|
tokenizers,
|
|
tomli,
|
|
tomli-w,
|
|
uv,
|
|
|
|
# tests
|
|
pytestCheckHook,
|
|
pytest-asyncio,
|
|
pytest-xdist,
|
|
qdrant-client,
|
|
vcrpy,
|
|
versionCheckHook,
|
|
writableTmpDirAsHomeHook,
|
|
}:
|
|
|
|
buildPythonPackage rec {
|
|
pname = "crewai";
|
|
version = "1.7.2";
|
|
pyproject = true;
|
|
|
|
src = fetchFromGitHub {
|
|
owner = "crewAIInc";
|
|
repo = "crewAI";
|
|
tag = version;
|
|
hash = "sha256-liJS3hrNsyAt49ROUi3/pWXfMP2wA/bFyy6yEOV6Rrk=";
|
|
};
|
|
|
|
sourceRoot = "${src.name}/lib/crewai";
|
|
|
|
build-system = [ hatchling ];
|
|
|
|
pythonRelaxDeps = [
|
|
"chromadb"
|
|
"click"
|
|
"json-repair"
|
|
"json5"
|
|
"litellm"
|
|
"mcp"
|
|
"openai"
|
|
"opentelemetry-api"
|
|
"opentelemetry-exporter-otlp-proto-http"
|
|
"opentelemetry-sdk"
|
|
"pdfplumber"
|
|
"portalocker"
|
|
"pydantic"
|
|
"pydantic-settings"
|
|
"pyjwt"
|
|
"python-dotenv"
|
|
"regex"
|
|
"tokenizers"
|
|
"tomli"
|
|
"tomli-w"
|
|
"uv"
|
|
];
|
|
|
|
dependencies = [
|
|
aiosqlite
|
|
appdirs
|
|
chromadb
|
|
click
|
|
instructor
|
|
json-repair
|
|
json5
|
|
jsonref
|
|
litellm
|
|
mcp
|
|
openai
|
|
opentelemetry-api
|
|
opentelemetry-exporter-otlp-proto-http
|
|
opentelemetry-sdk
|
|
openpyxl
|
|
pdfplumber
|
|
portalocker
|
|
pydantic
|
|
pydantic-settings
|
|
pyjwt
|
|
python-dotenv
|
|
regex
|
|
tokenizers
|
|
tomli
|
|
tomli-w
|
|
uv
|
|
];
|
|
|
|
pythonImportsCheck = [ "crewai" ];
|
|
|
|
disabledTestPaths = [
|
|
# Ignore tests that require {mem0, chromadb, telemetry, security, test_agent}
|
|
"tests/memory/test_external_memory.py" # require mem0ai
|
|
"tests/storage/test_mem0_storage.py" # require mem0ai
|
|
"tests/cli/test_git.py" # require git
|
|
"tests/memory/test_short_term_memory.py" # require require API keys
|
|
"tests/test_crew.py" # require require API keys
|
|
"tests/rag/chromadb/test_client.py" # issue with chromadb
|
|
"tests/telemetry/test_telemetry.py" # telemetry need network access
|
|
|
|
# ImportError: cannot import name 'InitFrom' from 'qdrant_client.models'
|
|
"tests/rag/qdrant/test_client.py"
|
|
|
|
# Tests requiring optional dependencies
|
|
"tests/llms/anthropic"
|
|
"tests/llms/azure"
|
|
"tests/llms/bedrock"
|
|
"tests/llms/google"
|
|
"tests/llms/litellm"
|
|
"tests/llms/hooks/test_anthropic_interceptor.py"
|
|
"tests/llms/hooks/test_unsupported_providers.py"
|
|
|
|
# Tests requiring network/API access
|
|
"tests/llms/openai"
|
|
"tests/telemetry"
|
|
"tests/tracing"
|
|
"tests/test_streaming_integration.py"
|
|
"tests/hooks"
|
|
"tests/llms/hooks/test_openai_interceptor.py"
|
|
"tests/test_llm.py"
|
|
|
|
# Tests requiring crewai-tools
|
|
"tests/agents/test_lite_agent.py"
|
|
];
|
|
|
|
disabledTests = [
|
|
# AssertionError - json-repair behaves differently since 0.52.0
|
|
"test_safe_repair_json_unrepairable"
|
|
"test_valid_action_parsing_with_curly_braces"
|
|
|
|
# Tests parser
|
|
"test_valid_action_parsing_with_special_characters"
|
|
|
|
# Tests agent - require API keys (OpenAI, Anthropic, etc)
|
|
"test_agent_custom_max_iterations"
|
|
"test_agent_error_on_parsing_tool"
|
|
"test_agent_execute_task_basic"
|
|
"test_agent_execute_task_with_context"
|
|
"test_agent_execute_task_with_custom_llm"
|
|
"test_agent_execute_task_with_ollama"
|
|
"test_agent_execute_task_with_tool"
|
|
"test_agent_execution"
|
|
"test_agent_execution_with_specific_tools"
|
|
"test_agent_execution_with_tools"
|
|
"test_agent_from_repository"
|
|
"test_agent_from_repository_override_attributes"
|
|
"test_agent_from_repository_with_invalid_tools"
|
|
"test_agent_function_calling_llm"
|
|
"test_agent_knowledege_with_crewai_knowledge"
|
|
"test_agent_moved_on_after_max_iterations"
|
|
"test_agent_powered_by_new_o_model_family_that_allows_skipping_tool"
|
|
"test_agent_powered_by_new_o_model_family_that_uses_tool"
|
|
"test_agent_remembers_output_format_after_using_tools_too_many_times"
|
|
"test_agent_repeated_tool_usage"
|
|
"test_agent_repeated_tool_usage_check_even_with_disabled_cache"
|
|
"test_agent_respect_the_max_rpm_set"
|
|
"test_agent_respect_the_max_rpm_set_over_crew_rpm"
|
|
"test_agent_step_callback"
|
|
"test_agent_use_specific_tasks_output_as_context"
|
|
"test_agent_with_knowledge_sources"
|
|
"test_agent_with_knowledge_with_no_crewai_knowledge"
|
|
"test_agent_with_ollama_llama3"
|
|
"test_agent_with_only_crewai_knowledge"
|
|
"test_agent_without_max_rpm_respects_crew_rpm"
|
|
"test_cache_hitting"
|
|
"test_custom_llm_temperature_preservation"
|
|
"test_custom_llm_with_langchain"
|
|
"test_disabling_cache_for_agent"
|
|
"test_do_not_allow_crewai_trigger_context_for_first_task_hierarchical"
|
|
"test_ensure_first_task_allow_crewai_trigger_context_is_false_does_not_inject"
|
|
"test_first_task_auto_inject_trigger"
|
|
"test_first_time_user_trace_collection_user_accepts"
|
|
"test_first_time_user_trace_collection_with_timeout"
|
|
"test_first_time_user_trace_consolidation_logic"
|
|
"test_get_knowledge_search_query"
|
|
"test_handle_context_length_exceeds_limit_cli_no"
|
|
"test_llm_call"
|
|
"test_llm_call_with_all_attributes"
|
|
"test_llm_call_with_ollama_llama3"
|
|
"test_logging_tool_usage"
|
|
"test_task_allow_crewai_trigger_context"
|
|
"test_task_allow_crewai_trigger_context_no_payload"
|
|
"test_task_without_allow_crewai_trigger_context"
|
|
"test_tool_result_as_answer_is_the_final_answer_for_the_agent"
|
|
"test_tool_usage_information_is_appended_to_agent"
|
|
"test_trace_batch_marked_as_failed_on_finalize_error"
|
|
|
|
# Tests lite agent - require API keys
|
|
"test_guardrail_is_called_using_callable"
|
|
"test_lite_agent_created_with_correct_parameters"
|
|
"test_guardrail_reached_attempt_limit"
|
|
"test_agent_output_when_guardrail_returns_base_model"
|
|
"test_lite_agent_with_tools"
|
|
"test_lite_agent_structured_output"
|
|
"test_lite_agent_returns_usage_metrics"
|
|
"test_guardrail_is_called_using_string"
|
|
|
|
# Tests evaluation - require API keys
|
|
"test_evaluate_current_iteration"
|
|
"test_eval_lite_agent"
|
|
"test_eval_specific_agents_from_crew"
|
|
"test_failed_evaluation"
|
|
|
|
# Tests knowledge - missing optional dependencies (docling, embedchain, pandas)
|
|
"test_docling_source"
|
|
"test_multiple_docling_sources"
|
|
"test_excel_knowledge_source"
|
|
|
|
# Test telemetry
|
|
"test_telemetry_fails_due_connect_timeout"
|
|
|
|
# Tests crew
|
|
"test_task_tools_override_agent_tools"
|
|
"test_crew_output_file_end_to_end"
|
|
"test_conditional_task_last_task_when_conditional_is_true"
|
|
"test_crew_with_failing_task_guardrails"
|
|
"test_kickoff_for_each_single_input"
|
|
"test_hierarchical_verbose_manager_agent"
|
|
"test_crew_function_calling_llm"
|
|
"test_replay_interpolates_inputs_properly"
|
|
"test_manager_agent_delegating_to_all_agents"
|
|
"test_crew_does_not_interpolate_without_inputs"
|
|
"test_crew_creation"
|
|
"test_delegation_is_not_enabled_if_there_are_only_one_agent"
|
|
"test_tools_with_custom_caching"
|
|
"test_api_calls_throttling"
|
|
"test_multimodal_agent_describing_image_successfully"
|
|
"test_warning_long_term_memory_without_entity_memory"
|
|
"test_replay_with_context"
|
|
"test_crew_verbose_output"
|
|
"test_before_kickoff_callback"
|
|
"test_hierarchical_verbose_false_manager_agent"
|
|
"test_task_with_no_arguments"
|
|
"test_replay_setup_context"
|
|
"test_kickoff_for_each_multiple_inputs"
|
|
"test_conditional_task_last_task_when_conditional_is_false"
|
|
"test_crew_with_delegating_agents"
|
|
"test_agents_do_not_get_delegation_tools_with_there_is_only_one_agent"
|
|
"test_multimodal_agent_live_image_analysis"
|
|
"test_hierarchical_process"
|
|
"test_crew_kickoff_usage_metrics"
|
|
"test_disabled_memory_using_contextual_memory"
|
|
"test_ensure_exchanged_messages_are_propagated_to_external_memory"
|
|
"test_agent_usage_metrics_are_captured_for_hierarchical_process"
|
|
"test_crew_log_file_output"
|
|
"test_before_kickoff_without_inputs"
|
|
"test_cache_hitting_between_agents"
|
|
"test_crew_kickoff_streaming_usage_metrics"
|
|
"test_async_crews_thread_safety"
|
|
"test_valid_action_parsing_with_unbalanced_quotes"
|
|
"test_valid_action_parsing_with_angle_brackets"
|
|
"test_valid_action_parsing_with_mixed_quotes"
|
|
"test_valid_action_parsing_with_parentheses"
|
|
"test_parsing_with_whitespace"
|
|
"test_valid_action_parsing_special_characters"
|
|
"test_parsing_with_special_characters"
|
|
"test_valid_action_parsing_with_newlines"
|
|
"test_valid_action_parsing_with_nested_quotes"
|
|
"test_valid_action_parsing_with_quotes"
|
|
"test_valid_action_parsing_with_escaped_characters"
|
|
"test_create_crew"
|
|
|
|
# Tests LLM - require API keys
|
|
"test_llm_callback_replacement"
|
|
"test_gemini_models"
|
|
"test_llm_call_with_message_list"
|
|
"test_gpt_4_1"
|
|
"test_o3_mini_reasoning_effort_low"
|
|
"test_handle_streaming_tool_calls"
|
|
"test_llm_call_with_string_input"
|
|
"test_llm_call_with_tool_and_string_input"
|
|
"test_llm_call_with_string_input_and_callbacks"
|
|
"test_llm_call_when_stop_is_unsupported"
|
|
"test_handle_streaming_tool_calls_with_error"
|
|
"test_o3_mini_reasoning_effort_medium"
|
|
"test_llm_call_with_tool_and_message_list"
|
|
"test_gemma3"
|
|
"test_llm_call_when_stop_is_unsupported_when_additional_drop_params_is_provided"
|
|
"test_o3_mini_reasoning_effort_high"
|
|
"test_handle_streaming_tool_calls_no_available_functions"
|
|
"test_handle_streaming_tool_calls_no_tools"
|
|
|
|
# Test main - require git
|
|
"test_publish_when_not_in_sync"
|
|
|
|
# Tests project - require embedchain and API keys
|
|
"test_before_kickoff_modification"
|
|
"test_before_kickoff_with_none_input"
|
|
"test_multiple_before_after_kickoff"
|
|
"test_after_kickoff_modification"
|
|
"test_internal_crew_with_mcp"
|
|
|
|
# Tests task - require API keys
|
|
"test_output_json_hierarchical"
|
|
"test_output_pydantic_sequential"
|
|
"test_no_inject_date"
|
|
"test_increment_tool_errors"
|
|
"test_task_execution_times"
|
|
"test_custom_converter_cls"
|
|
"test_save_task_output"
|
|
"test_inject_date"
|
|
"test_output_pydantic_hierarchical"
|
|
"test_json_property_without_output_json"
|
|
"test_task_interpolation_with_hyphens"
|
|
"test_increment_delegations_for_hierarchical_process"
|
|
"test_save_task_json_output"
|
|
"test_output_pydantic_to_another_task"
|
|
"test_output_json_dict_sequential"
|
|
"test_inject_date_custom_format"
|
|
"test_output_json_sequential"
|
|
"test_output_json_to_another_task"
|
|
"test_increment_delegations_for_sequential_process"
|
|
"test_save_task_pydantic_output"
|
|
"test_output_json_dict_hierarchical"
|
|
"test_task_with_max_execution_time"
|
|
"test_task_with_max_execution_time_exceeded"
|
|
|
|
# Tests task guardrails
|
|
"test_task_guardrail_process_output"
|
|
"test_guardrail_when_an_error_occurs"
|
|
"test_guardrail_emits_events"
|
|
|
|
# Tests tools
|
|
"test_delegate_work"
|
|
"test_delegate_work_with_wrong_co_worker_variable"
|
|
"test_ask_question_with_wrong_co_worker_variable"
|
|
"test_ask_question"
|
|
"test_delegate_work_withwith_coworker_as_array"
|
|
"test_ask_question_with_coworker_as_array"
|
|
"test_max_usage_count_is_respected"
|
|
"test_async_tool_using_within_isolated_crew"
|
|
"test_async_tool_using_decorator_within_flow"
|
|
"test_async_tool_using_decorator_within_isolated_crew"
|
|
"test_async_tool_within_flow"
|
|
"test_ainvoke"
|
|
|
|
# Tests tracing
|
|
"test_trace_listener_disabled_when_env_false"
|
|
"test_trace_listener_with_authenticated_user"
|
|
"test_trace_listener_collects_crew_events"
|
|
"test_trace_listener_ephemeral_batch"
|
|
"test_events_collection_batch_manager"
|
|
|
|
# Tests utilities
|
|
"test_converter_with_llama3_2_model"
|
|
"test_converter_with_llama3_1_model"
|
|
"test_converter_with_nested_model"
|
|
"test_convert_with_instructions"
|
|
"test_crew_emits_start_kickoff_event"
|
|
"test_crew_emits_test_kickoff_type_event"
|
|
"test_crew_emits_start_task_event"
|
|
"test_llm_emits_call_started_event"
|
|
"test_register_handler_adds_new_handler"
|
|
"test_multiple_handlers_for_same_event"
|
|
"test_llm_emits_event_with_lite_agent"
|
|
"test_tools_emits_finished_events"
|
|
"test_agent_emits_execution_started_and_completed_events"
|
|
"test_crew_emits_end_kickoff_event"
|
|
"test_llm_emits_stream_chunk_events"
|
|
"test_tools_emits_error_events"
|
|
"test_crew_emits_end_task_event"
|
|
"test_stream_llm_emits_event_with_task_and_agent_info"
|
|
"test_llm_no_stream_chunks_when_streaming_disabled"
|
|
"test_llm_emits_event_with_task_and_agent_info"
|
|
|
|
# Tests qdrant
|
|
"test_acreate_collection"
|
|
"test_aget_or_create_collection_wrong_client_type"
|
|
"test_aadd_documents_empty_list"
|
|
"test_acreate_collection_already_exists"
|
|
"test_aget_or_create_collection_existing"
|
|
"test_aadd_documents_collection_not_exists"
|
|
"test_acreate_collection_wrong_client_type"
|
|
"test_aadd_documents"
|
|
"test_aadd_documents_wrong_client_type"
|
|
"test_asearch"
|
|
"test_adelete_collection_not_exists"
|
|
"test_asearch_with_filters"
|
|
"test_adelete_collection_wrong_client_type"
|
|
"test_asearch_collection_not_exists"
|
|
"test_areset"
|
|
"test_areset_no_collections"
|
|
"test_aget_or_create_collection_new"
|
|
"test_aadd_documents_with_doc_id"
|
|
"test_asearch_wrong_client_type"
|
|
"test_areset_wrong_client_type"
|
|
"test_adelete_collection"
|
|
|
|
# Tests requiring litellm
|
|
"test_litellm_anthropic_error_handling"
|
|
"test_litellm_auth_error_handling"
|
|
"test_crew_agent_executor_litellm_auth_error"
|
|
"test_agent_with_callbacks"
|
|
"test_crew_memoization"
|
|
"test_task_guardrail"
|
|
"test_crew_name"
|
|
"test_get_conversion_instructions_non_gpt"
|
|
"test_supports_function_calling_false"
|
|
|
|
# Tests requiring network
|
|
"test_agent_max_iterations_stops_loop"
|
|
"test_task_output_includes_messages"
|
|
|
|
# Not work with nixpkgs-review
|
|
"test_concurrent_ainvoke_calls"
|
|
];
|
|
|
|
nativeCheckInputs = [
|
|
pytestCheckHook
|
|
pytest-asyncio
|
|
pytest-xdist
|
|
qdrant-client
|
|
vcrpy
|
|
versionCheckHook
|
|
writableTmpDirAsHomeHook
|
|
];
|
|
|
|
pytestFlagsArray = [
|
|
"--override-ini=addopts="
|
|
];
|
|
|
|
meta = {
|
|
description = "Framework for orchestrating role-playing, autonomous AI agents";
|
|
homepage = "https://github.com/crewAIInc/crewAI";
|
|
changelog = "https://github.com/crewAIInc/crewAI/releases/tag/${version}";
|
|
license = lib.licenses.mit;
|
|
maintainers = with lib.maintainers; [ liberodark ];
|
|
platforms = lib.platforms.linux;
|
|
mainProgram = "crewai";
|
|
};
|
|
}
|